Comparing and combining GLM and SVM

This notebook accompanies below paper:

F. Sepehrband\*, K.M. Lynch, R.P. Cabeen, C. Gonzalez-Zacarias, L. Zhao, M. D'Arcy, C. Kesselman, M.M. Herting, I.D. Dinov, A.W. Toga, K.A. Clark, **Neuroanatomical Morphometric Characterization of Sex Differences in Youth Using Multivariate Statistical Learning**, *NeuroImage*, submitted September 2017.

In this notebook, we compare results of GLM with SVM. Then, we combine them in a single figure, with the aim of increased interpretability.

  • Strucutral MRI of the PNC study were used for this study.
  • Morphological features of the strucutral images were derived using FreeSurfer toolkit.
  • FreeSurfer analysis was performed using LONI pipeline on high performance computing of **USC** Mark and Mary Stevens Neuroimaging and Informatics Institute, Keck school of Medicine of **USC**. The outputs were QC'ed and used as the input of this notebook.

author:
Farshid Sepehrband,
Laboratory of Neuro Imaging, USC Mark and Mary Stevens Neuroimaging and Informatics Institute, Keck School of Medicine of USC, University of Southern California, Los Angeles, CA, USA

farshid.sepehrband@loni.usc.edu
@fsepehrband

Importing required libraries

In [1]:
import numpy as np
import matplotlib.pyplot as plt
import plotly.plotly as py
import pandas as pd 
import os
%pylab inline
Populating the interactive namespace from numpy and matplotlib

Comparing GLM results with SVM

Below are maps of statistical measures derived from GLM and SVM. t-statistics of GLM and beta coefficients of SVM are plotted.

In [2]:
from IPython.display import Image
Image(filename='../files/demoFig.png')
Out[2]:

Below video compares these maps.

In [3]:
from IPython.display import YouTubeVideo
vid = YouTubeVideo("QjyAV9QPiXU")
display(vid)

Combining SVM beta, correlation with brain size and GLM statistics

Below figure combines SVM beta, GLM t-statistic and correlation with brain size.

Interactive plot is presented here.

Follow the instruction below to work with the interactive plot:

  • Click on the link above,
  • Close the pop up window (or sign up),
  • At top right corner of the plot, click on alt text
  • Hover over bubbles to see statistical information
In [4]:
Image(filename='../files/bubble.jpg')
Out[4]:
In [ ]: